fix: sanitize 502 HTML responses from error messages

Hellnight2005 · Hellnight2005 · commit d887401cc983 · 2026-02-04T16:24:59.000+05:30
This commit fixes an issue where 502 Bad Gateway responses from upstream proxies (returning raw HTML) would cause unreadable exceptions. The error handling now attempts to parse JSON and falls back to a clean status message if the response body is HTML. Includes unit tests.
diff --git a/src/lingodotdev/engine.py b/src/lingodotdev/engine.py
@@ -185,9 +185,17 @@ async def _localize_chunk(
 
             if not response.is_success:
                 if 500 <= response.status_code < 600:
+                    error_details = ""
+                    try:
+                        error_json = response.json()
+                        if isinstance(error_json, dict) and "error" in error_json:
+                            error_details = f" {error_json['error']}"
+                    except Exception:
+                        pass
+
                     raise RuntimeError(
-                        f"Server error ({response.status_code}): {response.reason_phrase}. "
-                        f"{response.text}. This may be due to temporary service issues."
+                        f"Server error ({response.status_code}): {response.reason_phrase}.{error_details} "
+                        "This may be due to temporary service issues."
                     )
                 elif response.status_code == 400:
                     raise ValueError(
@@ -427,8 +435,16 @@ async def recognize_locale(self, text: str) -> str:
 
             if not response.is_success:
                 if 500 <= response.status_code < 600:
+                    error_details = ""
+                    try:
+                        error_json = response.json()
+                        if isinstance(error_json, dict) and "error" in error_json:
+                            error_details = f" {error_json['error']}"
+                    except Exception:
+                        pass
+
                     raise RuntimeError(
-                        f"Server error ({response.status_code}): {response.reason_phrase}. "
+                        f"Server error ({response.status_code}): {response.reason_phrase}.{error_details} "
                         "This may be due to temporary service issues."
                     )
                 raise RuntimeError(
@@ -461,8 +477,16 @@ async def whoami(self) -> Optional[Dict[str, str]]:
                     return {"email": payload["email"], "id": payload["id"]}
 
             if 500 <= response.status_code < 600:
+                error_details = ""
+                try:
+                    error_json = response.json()
+                    if isinstance(error_json, dict) and "error" in error_json:
+                        error_details = f" {error_json['error']}"
+                except Exception:
+                    pass
+
                 raise RuntimeError(
-                    f"Server error ({response.status_code}): {response.reason_phrase}. "
+                    f"Server error ({response.status_code}): {response.reason_phrase}.{error_details} "
                     "This may be due to temporary service issues."
                 )
 
diff --git a/tests/test_502_handling.py b/tests/test_502_handling.py
@@ -0,0 +1,60 @@
+import pytest
+from unittest.mock import Mock, patch
+from lingodotdev import LingoDotDevEngine
+
+
+@pytest.mark.asyncio
+async def test_502_html_handling():
+    """Test that 502 errors with HTML bodies are sanitized"""
+    config = {"api_key": "test_key", "api_url": "https://api.test.com"}
+    engine = LingoDotDevEngine(config)
+
+    html_body = """<html><body><h1>502 Bad Gateway</h1></body></html>"""
+
+    with patch("lingodotdev.engine.httpx.AsyncClient.post") as mock_post:
+        mock_response = Mock()
+        mock_response.is_success = False
+        mock_response.status_code = 502
+        mock_response.reason_phrase = "Bad Gateway"
+        mock_response.text = html_body
+        mock_response.json.side_effect = ValueError(
+            "Not JSON"
+        )  # simulating non-JSON response
+        mock_post.return_value = mock_response
+
+        with pytest.raises(RuntimeError) as exc_info:
+            await engine.localize_text("hello", {"target_locale": "es"})
+
+        error_msg = str(exc_info.value)
+
+        # Assertions
+        assert "Server error (502): Bad Gateway." in error_msg
+        assert "This may be due to temporary service issues." in error_msg
+        assert "<html>" not in error_msg
+        assert "<body>" not in error_msg
+
+
+@pytest.mark.asyncio
+async def test_500_json_handling():
+    """Test that 500 errors with JSON bodies are preserved"""
+    config = {"api_key": "test_key", "api_url": "https://api.test.com"}
+    engine = LingoDotDevEngine(config)
+
+    error_json = {"error": "Specific internal error message"}
+
+    with patch("lingodotdev.engine.httpx.AsyncClient.post") as mock_post:
+        mock_response = Mock()
+        mock_response.is_success = False
+        mock_response.status_code = 500
+        mock_response.reason_phrase = "Internal Server Error"
+        mock_response.json.return_value = error_json
+        mock_post.return_value = mock_response
+
+        with pytest.raises(RuntimeError) as exc_info:
+            await engine.localize_text("hello", {"target_locale": "es"})
+
+        error_msg = str(exc_info.value)
+
+        # Assertions
+        assert "Server error (500): Internal Server Error." in error_msg
+        assert "Specific internal error message" in error_msg
